Re: Indian scripts and Unicode

[Leong Kok Yong <ccelky@valentine.irdu.nus.sg>]

hi Jeroen,

Yes, you're right. Unicode may solved some problems but it created a very
huge problem especially when it comes to implementation. You need to write
helper applications to run alongside your web browser in order to see
these Tamil characters in Unicode. You can do that for Windows and Mac,
but definitely not UNIX to do rendering. And such expensive software
solutions usually becomes commercial software due to the huge resources
put into its development. The masses will not benefit.


But with a 8-bit codeset, you can just tell your browser to simply switch
the font. Most browser can do that nowadays. Fonts are easily distriubted
free for most platforms from Windows, Mac to various flavours of UNIX. Not
only that, many free keyboard managers can be easily configured to work
with 8-bit font code to allow users to type. Do you know of any FREE
application that can allow you to type Tamil Unicode characters on most
platforms? 


I'm glad that you understand the current situation that there is too many 
7-bit and 8-bit Tamil fontset around. That's exactly why those developers 
are coming together to standardize on a universal one, which hopefully 
will be adopted by all involved.

As for Tamil numerals, i'll leave it to the experts on the mailing 
lists to help answer your questions.... 




On Tue, 9 Sep 1997, Jeroen Hellingman wrote:

> Date: Tue, 9 Sep 1997 08:59:10 +0200
> From: Jeroen Hellingman <etmjehe@genesis.etm.ericsson.se>
> To: ccelky@valentine.irdu.nus.sg
> Subject: Re: Indian scripts and Unicode
> 
> 
> Hi,
> 
> It is relatively easy to create a standard Tamil standard, containing all
> glyphs (that is shapes of letters and combinations of letters) as
> separate entities. In Tamil this is relatively straightforward, as
> there are no conjuncts (except ksa), and the number of letters is much
> lower than
> for other Indian languages, and it will be very easy for presentation
> purposes---but still I don't think this is the best approach. All other
> processing (like sorting, searching, and so on) will be far more
> complicated. The approach of ISCII has been to encode logical characters,
> that is, the basic alphabet and vowel signs, and then have the
> software select the appropriate glyphs on the fly. This makes 
> presentation a bit more difficult (and impossible on applications
> that use the narrowminded view that there is a one-to-one mapping
> from characters to glyphs) but all other processing far more convenient.
> 
> Have a look at the current situation---about every font for Tamil
> uses its own organisation. Some have the Roman alphabet as well, some
> not, and every single font has its Tamil glyphs in different positions,
> making Tamil on the web very font-dependent. Creating a standard
> for this now will not solve the problem overnight, but only add
> yet another Tamil font encoding. The approach used in ISCII and Unicode
> is to seperate the issue of characters from glyphs, and encode the
> glyphs. Then it is relatively easy to do a translation from characters
> to glyphs, using a mapping table associated with the font. (which, if
> best is incorporated in the font, so changing to a font with a 
> different encoding will be painless for the user). This is the 
> approach I used when making script tools for Indian scripts. If
> you are looking for a reasonable standard for Tamil encoding, have
> a look at the fonts made by C-DAC. If you wish, I can give the
> encoding used in that one (that is, inferred from a font.) For other
> Indian scripts, like Devanagari or Malayalam, the situation is
> much more complex, as the number of required glyphs easily exceeds
> the number of glyphs you have available in an 8-bit font (about 230),
> so even more tricks are required.
> 
> Do you speak Tamil? I am doing some research to the use of Tamil 
> numerals. The Unicode standard shows the old, non-decimal numbers,
> but on a banknote from Mauritius, I also saw some other numerals,
> including a zero, which are supposed to be Tamil. Do have any idea
> about these?
> 
> I will probably do a font for Tamil after I complete my Oriya font,
> and have done a clean-up on my Malayalam font. It will include such
> things as wide versions of pa to do line adjustments (sometimes seen
> in handwriting)
> 
> Best Regards,
> 
> Jeroen
> +-----------------------------------+-----------------------------+
> | Jeroen Hellingman                 | private:                    |
> | Ericsson Telecommunicatie B.V.    | 't Zand 2                   |
> | Ericssonstraat 2, Rijen, Holland  | 4133 TB Vianen              |
> | Department ETM/RPU, Room 17107    | The Netherlands             |
> | Tel: +31 161 242022   (834 2022)  |                             |
> | Fax: +31 161 249699               | Tel: +31 347 373935         |
> | E-mail: <etmjehe@etm.ericsson.se> |                             |
> +-----------------------------------+-----------------------------+
>